Paul Graham, co-founder of Y Combinator, explains how a startup can make its founders billionaires without cheating. Speaking at the Oxford Union (June 2026), he draws on 21 years of experience: YC, which he started with Jessica Livingston, is a cross between an investment firm and a school for startup founders, and since 2005 it has funded about 6500 companies. Starting a successful startup is the most common route to becoming a billionaire, so Graham has effectively spent 21 years training people to become billionaires — roughly 30 so far, with many more in the pipeline.
The central claim is that becoming a billionaire does not require cheating. Graham was astonished when an American politician said it was impossible to earn a billion dollars without doing something bad — he compares it to a skating coach hearing that a triple axel is impossible. The key to getting that rich is exponential growth, which is determined by only two numbers: the monthly growth rate and how long that growth continues. He demonstrates that starting from modest revenue, even a 15% monthly growth rate over five years leads to a valuation that makes the founder a billionaire – provided the market is large enough. You get the growth rate by making something users like so much they tell their friends, and the duration by being in a big market. Grow exponentially into a big market and the startup becomes valuable, and you, as a shareholder, become rich — automatically, just by keeping customers happy. For young founders, the best ideas come from making things they themselves want, often while working on projects with friends rather than deliberately searching for startup ideas.
Key Points
Y Combinator, founded in 2005 (21 years ago) by Paul Graham and Jessica Livingston, is a cross between an investment firm and a school for startup founders; it has funded about 6500 companies and produced roughly 30 billionaires so far, with many more in the pipeline.
Exponential growth makes billionaires: a startup growing 93% per month from $2M reaches billionaire status in about 9.45 months.
A 15% monthly growth rate is not rare — Graham constantly encounters it; after five years revenue multiplies roughly 4,384 times.
Only two numbers determine how rich a founder becomes: the growth rate and how long it continues.
The 93% example: a founder Graham had funded reported 93% growth last month. Someone online objected that a few million growing at 93%/month is "radically different" from being a billionaire. Graham says this is false in a very illuminating way.
The math: take "a few" conservatively as 2 million, so the company must grow 500x. Months of 93% growth needed = log base 1.93 of 500 ≈ 9.45 (just type log(500, 1.93) into Google). A couple of million and a billion are nine and a half months apart.
The 15% arithmetic: 1.15^60 ≈ 4384 over five years (60 months). At $10,000/month now, that's about $44 million/month in five years, or $526 million a year — at which point, owning what founders typically own, you'd be a billionaire.
In reality growth rates slow a bit: a very successful startup probably grows faster than 15%/month in year 1 and slower in year 5, but you end up in about the same place. Starting in your early twenties, becoming a billionaire by thirty is hard but possible.
Exponential growth is like magic, generating outcomes that seem impossible — which is why some politicians distrust it. Not understanding the math, they assume cheating when they see seemingly impossibly rich people.
Cheating cannot affect market size: to grow 4000x there must be at least 4000x more demand, so any sustainable growth is honest.
The best startup ideas emerge unconsciously from building what seems cool, not from active searching.
Apple, Google, Facebook, and Twitch all started as non-commercial projects.
The key to exponential growth is deep user empathy – understanding what users want and making their lives dramatically better.
How a society produces rich people is one of the most important things to understand about it. Beliefs about this should not come from ideology, movies, or centuries-old historical examples — you must look at the world around you and see how it is actually done. Future prime ministers especially need to remember this.
Concepts
Exponential Growth: A constant percentage increase each month; even small rates produce enormous outcomes over time, generating results that look impossible.
Growth Rate: The monthly percentage increase in revenue or users (and thus the founder's stake), driven by word-of-mouth from a product users love. It is the first thing Graham asks a founder, and it serves as a diagnostic: to sustain a consistent monthly rate, you must make something so good that people tell their friends, so the number reveals whether you've built the right thing.
Duration: How long exponential growth continues, bounded by market size. Growing 4000x requires at least 4000x more demand; growth without cheating is limited by how big the market can become.
Word of Mouth: The non-cheating mechanism behind consistent growth. Users love what you built, you exploit no one, you work hard to make users happy, and users tell their friends, producing exponential growth.
Beachhead Market: A small initial user group (e.g., oneself or a niche) that can expand as the need spreads. Since you predict future demand, the market will grow, and you can always expand into adjacent markets.
User Empathy: The ability to understand a group of users so well that one can create exactly what they want – the engine of organic growth.
Unsatisfiable-Need Discovery: The problem (and great thing) about market economies is that it's hard to make something customers want that they don't already have; as soon as a new, satisfiable need is discovered, people rush to satisfy it. So you must discover a need no one else knows about yet — by feeling the need yourself.
The Second Signal (make what you and your friends want): Young founders usually lack the experience to know what others need, so they should make something they themselves want. Their own needs are uniquely valuable because they predict future demand: you're the age when people start using new things, and whatever you and your friends start using now, everyone will use in ten years. Intuitions about other people's needs are usually a crap signal; your own needs are an especially valuable one. It needn't be a consumer product — maybe you and your friends are molecular biologists and there's something cool to be done now to DNA that everyone else overlooked, or you're into drones. The idea needn't have wide appeal; literally just appeal to you and your friends.
Unconscious Idea Generation / Not Looking for Startup Ideas: One of the most counterintuitive things about startups is that the way to get the very best startup ideas is not to look for startup ideas. Conscious searching makes you too conservative; you'll lop off the outliers. The phenomenon is that the unconscious mind recognizes good startup ideas before the conscious mind does, so building something that genuinely seems cool often leads to a viable business.
Details
Paul Graham, speaking at the Oxford Union, argues that the belief that one cannot become a billionaire without cheating is mathematically false. He frames the whole talk around a calculation he does constantly as an investor. When a founder he'd funded told him she was growing 93% last month — meaning her net worth was also growing 93% a month — he noted she had done nothing bad. Users loved what she'd built, she wasn't exploiting anyone, and she and her cofounder had worked hard to make users happy, so users told their friends; that produced the exponential growth. A critic replied that having a few million growing at 93% a month is radically different from being a billionaire. Graham treats this as not merely false but false in a very illuminating way: interpreting "a few" most conservatively as 2 million, the company must grow 500x, and months of 93% growth needed = log base 1.93 of 500 ≈ 9.45. A couple of million and a billion are therefore nine and a half months apart — not radically different at all. The calculation involves only two numbers – growth rate and time – so if becoming a billionaire is impossible without cheating, either the growth rate or the time must be impossible. This is why the first question he asks a founder is their growth rate.
He then runs the same math at a more conservative 15% a month. Over five years (60 months) at 15%, revenue multiplies by about 4,384 (1.15^60). A startup earning $10,000 per month would then generate $44 million per month ($526 million per year). At that scale, the founder's typical equity stake makes them a billionaire. Real trajectories vary — faster than 15%/month in year 1, slower in year 5 — but land in about the same place. A successful startup founded in one's early twenties can still make its founder a billionaire by age thirty. The purpose of doing the math yourself is that exponential growth behaves like magic, generating outcomes that look impossible, which is exactly why some politicians distrust it. Lacking the math, they see people becoming seemingly impossibly rich and assume cheating.
The duration of growth depends on market size; to grow 4000x there must be at least 4000x more demand, and market size cannot be cheated upward, so honest growth is the only viable path. To sustain a consistent monthly rate, the startup must make something so good that users spontaneously recruit new users. Graham therefore asks every YC startup about their growth rate as a diagnostic: if it's low, the product isn't good enough yet. Market economies make this hard (and great): it's hard to make something customers want that they don't already have, and the moment a new satisfiable need appears, people rush to fill it. The winning move is to discover a need nobody else knows about yet by feeling it yourself.
For young founders, the most reliable way to discover such a product is to build something they themselves want. Their own needs predict future demand because whatever they and their friends start using now will be used by everyone in ten years — whereas their guesses about other people's needs are usually a crap signal. This need not be a consumer product – it could be a tool for molecular biologists or drone enthusiasts. The idea needn't have broad appeal, and market size should not be a worry; all you need is a beachhead in the territory of unsatisfied need, and the market will grow into your prediction, with adjacent markets always available for expansion. Graham emphasizes that the best startup ideas come not from deliberate search but from working on projects with friends that seem cool, not intended as companies. Conscious searching makes you too conservative and lops off the outliers. Apple, Google, and Facebook all began this way. Justin.TV, a project that seemed frivolous, eventually pivoted into Twitch, a billion-dollar business. The unconscious mind recognizes promising ideas before the conscious mind does: anything that genuinely feels like a cool thing to build has a high probability of being a good startup idea, no matter how preposterous it sounds. (The notes of the talk break off mid-thought here: "The best ideas sound…")
The path to exponential growth is not exploitation but empathy. The founder must understand a group of users so well that they can create exactly what those users want. For young founders, the hack is to make something for themselves. This approach automatically ensures deep empathy and a product that people will love.
Finally, on the societal level: how people become rich in a society is among the most important things to understand about it, and beliefs about it should be formed by looking at the world around you, not by ideology, movies, or centuries-old examples. Anyone who wants to become rich will be forced to understand the mechanism anyway. The people who most need to remember this are future prime ministers.
Rendering diagram…
Graham thanks Trevor Blackwell, Jared Friedman, Jessica Livingston, and Garry Tan for reading drafts, and Arwa Elrayess and the Oxford Union for hosting.